Help understanding a schematic and transistor question

[jfrabat]

BACKGROUND:

My daughter finally got a computer in her room (now that she is old enough), and she plays piano and sings.  She wants to be able to record in her own room, so she saved up and bought a Presonus interface for her PC, and she has 2 MIDI keyboards (she also has a USB Mic, but, without the interface, the lag is horrendous).  Now that she is getting a bit more involved (and because she has experience with my gear!), she has been asking me to build her some stuff.  I made her an EQ (well, almost done; jsut waiting for some pots to arrive), and I started on a pre-amp also.  But she asked me for a compressor (her PC is my wife's old PC, so plugins make the lag bad).

So, looking around for options, an old (1985) article of Electronics Today was sent my way, which has a Combined Direct Inject Compressor and Noise Gate, which I really like (it has some nice features, and can be used for Mic or instrument - or even amplified signals, for that matter).

THE COMPRESSOR:

This is the link I was sent (you can reed the complete article in pages46-52).  This is the Schematic:



And the corresponding parts list:



WHY I LIKE IT:

There are many things I like about this one.  In fact, if it had phantom power, it would check all the boxes (and even that is not that important because the preamp I am making will have it).  Things that I like are:

• Input options; Mic (XLR and TRS), instrument (2 of them), and even amplified signals
• Side chain for the compressor
• External trigger for the gate
• Depth adjustment on the gate
• The project is challenging (would be my most complicated project to date) but not impossible
• It is kind of cool to finally build something off a magazine (as a teenager in the 80's, I would read all the RC Model magazines, but never did build anything; as an adult, I contributed to magazines and had a couple or articles published.  Now I would finally be BUILDING off a magazine; it is just one of those things I never checked off the list!)

INTENDED MODS:

I do plan to slightly modify the schematic in the following ways:

• Get rid of the push/pull pots and just add regular switches for selecting to bypass the noise gate and the -10 dB option
• Use a combo jack for the mic in instead of an XLR + TRS
• Replace PR4 with a regular pot to have easy access to the compressor attack
• Replace PR9 (incorrectly labeled RV9 in the parts list) with a pot to have easy access to the noise gate depth

So very minor mods, as you can see...

QUESTIONS:

Now, regarding the schematic, I have a couple of questions:

• Where the heck is Switch 2?  SW1 is clear (violet), and that one controls the -10 dB sensitivity.  But SW2 is supposed to bypass the gate, yet I cannot see it in the schematic.  Unless it is tied to the red, which is my next question...  (my guess is that it is, but can you confirm?)
• What the heck is the red?  Is it supposed to be the SW2 switching VE+ and VE- for IC5?  I am still not clear what they mean...
• What the heck does the symbols in orange and blue mean?
• What kind of plug is SK9 and where do I get it?

In addition, I have the following questions not related to the schematic:

• Where is a good place to get a power source with +/- 15V?  I could use a charge pump like I do in the EQ, but if possible, I would rather just use a brick outside of the circuit.  Anyone know where I can order these online?  And also, the corresponding plug, as stated above...
• What is a good replacement for BC212L?  Seems those are a bit harder to get than they were 35 years ago!  Neither Mouser nor Tayda carries them!  Would 2N3906 be a suitable replacement?

Thanks in advanced for your help, guys!  I am really looking forward to this project!

[kraal]

Hi,

What the heck does the symbols in orange and blue mean?

These are wave forms. Shown on the schematic to better understand what's happening where they are shown.

What is a good replacement for BC212L?

Use a NTE replacement: NTE159 [1]
Just check the prices at different distributors: it's 66 ct/pc at Arrow Electronics, but 5CHF/pc at grieder bauteile (8x the unit price of Arrow)

[1] http://dilp.netcomponents.com/cgi-bin/nteinc.asp?partnumber1=NTE159

[jfrabat]

OK, I thought that's what they were.  So what about the reds and SW2?  Anyone wanna take a stab at those?

By the way, I also noticed an error; in the parts list, IC4 is a TL074 (also stated in the schematics).  But in IC4c, they list pin 9 as IN+ and 10 as IN-, which is incorrect.  Looks like it is only the pin labeling that is wrong, though (they accidentally got them backwards - the price of not doing them in a computer, I guess!  After all, they WERE made in 1985!).

[Rob Strand]

OK, I thought that's what they were.  So what about the reds and SW2?  Anyone wanna take a stab at those?

The red stuff are just waveforms at those points in the circuit.  The idea is to let you understand the circuit better:
- bumps = half-wave rectified sine wave
- steps = signals going on/off
-  hill = envelope of signal on the cap


All the RV6, SW2 stuff is confusing.     I can't make everything out from the traces in figure 9.
You will need to trace the board.

Figure 11 shows the wiring for SW1, SW2, RV5, RV6.

To me it is odd that SW1 which seems to be related to the external trigger but it appears on RV5 (release) and not RV6 (external trigger level).

For more info look at fig 9:
- Wires on the PCB are marked with little arrows.
   At the top the two arrows appear to be the connections for SW1.

- Near  RV6 and R70 you can see arrows with b and c.   
  At first I thought it was just the connections to the pot RV6!

- Possibly under IC4 and near R53, R59 is the connection for 'a'.   I can't quite make it out.


Not so easy to work out these last two since the PCB only shows the tracks for one side of the board.   


More detailed pics of the PCB traces might appear elsewhere in the magazine.   Tracing the PCB might help make sense of the 'a', 'b' and 'c' connections.

By the way, I also noticed an error; in the parts list, IC4 is a TL074 (also stated in the schematics).  But in IC4c, they list pin 9 as IN+ and 10 as IN-, which is incorrect.  Looks like it is only the pin labeling that is wrong, though (they accidentally got them backwards - the price of not doing them in a computer, I guess!  After all, they WERE made in 1985!).

In April 1987 issue there's an index for projects.   In the index they show what issues have the "errata" for each project.  I could not see any errata for this project.   Not for the things you mentioned and not for RV6 and SW2.

Q3 is not labelled on the schematic.    Probably the transistor at the top right.

[jfrabat]

Ron,

So a couple of questions:

1. You think it is worth it trying this schematic even though the SW2 situation is not clear?  I have been passing transcribing it to Eagle...
2. How would you go about fixing the noise gate bypass? 
3. Which replacement would you recommend for the BC212?

OK, I thought that's what they were.  So what about the reds and SW2?  Anyone wanna take a stab at those?

The red stuff are just waveforms at those points in the circuit.  The idea is to let you understand the circuit better:
- bumps = half-wave rectified sine wave
- steps = signals going on/off
-  hill = envelope of signal on the cap

OK, got it!

All the RV6, SW2 stuff is confusing.     I can't make everything out from the traces in figure 9.
You will need to trace the board.

Figure 11 shows the wiring for SW1, SW2, RV5, RV6.

To me it is odd that SW1 which seems to be related to the external trigger but it appears on RV5 (release) and not RV6 (external trigger level).

According to the article (page 46, middle column, last paragraph), the purpose of this switch is to reduce sensitivity by 10 dB.  So that makes a bit more sense to have it in the release.  Although, I cannot figure out how that switch achieves that.  Granted, I don't understand about 3 quarters of the schematic also!

Here is the board image, for the convenience of those reading in:

For more info look at fig 9:
- Wires on the PCB are marked with little arrows.
   At the top the two arrows appear to be the connections for SW1.

- Near  RV6 and R70 you can see arrows with b and c.   
  At first I thought it was just the connections to the pot RV6!

- Possibly under IC4 and near R53, R59 is the connection for 'a'.   I can't quite make it out.


Not so easy to work out these last two since the PCB only shows the tracks for one side of the board.   

More detailed pics of the PCB traces might appear elsewhere in the magazine.   Tracing the PCB might help make sense of the 'a', 'b' and 'c' connections.

By the way, I also noticed an error; in the parts list, IC4 is a TL074 (also stated in the schematics).  But in IC4c, they list pin 9 as IN+ and 10 as IN-, which is incorrect.  Looks like it is only the pin labeling that is wrong, though (they accidentally got them backwards - the price of not doing them in a computer, I guess!  After all, they WERE made in 1985!).

In April 1987 issue there's an index for projects.   In the index they show what issues have the "errata" for each project.  I could not see any errata for this project.   Not for the things you mentioned and not for RV6 and SW2.

Q3 is not labelled on the schematic.    Probably the transistor at the top right.

Yeah, I noticed Q3 when re-drawing the schematic in Eagle.  Since there is no Q3 in the schematic, and considering it is next to Q2 (which is next to Q1) and that the bottom has Q4~6, I am certain that the transistor on the top right is Q3.

I also noticed that IC5c and IC6c have Pin9 as - and 10 as + (it should be backwards, right?  Positive is 9 and negative is 10 according to all the pinouts I have seen online).

Regarding the board, I marked the switches and outputs here:



I agree with you that the top 2 must be SW1.  And I think you are correct about A (and B and C).  I also marked (red) VE+, VE- and GND.  And for those reading in, to save you from the confusion I had when I first read it, the SW in the external trigger (yellow) is from the jack.

I will try to trace SW2 to see what it is connecting to.  Will post again once I have done it.

[ElectricDruid]

Replace PR4 with a regular pot to have easy access to the compressor attack

It looks to me like PR4 and RV5 just set the attack/decay time constants for the LED level display. They're nothing to do with the compressor rectifier, which is all down in the bottom-left corner, not the top-right. (I think...)

[jfrabat]

Replace PR4 with a regular pot to have easy access to the compressor attack

It looks to me like PR4 and RV5 just set the attack/decay time constants for the LED level display. They're nothing to do with the compressor rectifier, which is all down in the bottom-left corner, not the top-right. (I think...)

I think they adjust everything in the compressor.  RV 5 is a pot already (in the original).  As for PR4, read the highlighted part...

[Rob Strand]

I had a chance to go over things again.

According to the article (page 46, middle column, last paragraph), the purpose of this switch is to reduce sensitivity by 10 dB.  So that makes a bit more sense to have it in the release.  Although, I cannot figure out how that switch achieves that.  Granted, I don't understand about 3 quarters of the schematic also!

OK that's useful!

So now it does make sense for SW1 to be on RV5.   When SW1 is closed is drops the gain of IC2a.

I'm now fairly convinced there's no SW2.  The connections for SW2 are actually wired connections for RV6. 
The reason for that are:
- RV6 has no pads on the PCB.
- The 'a' 'b' 'c' points seem to wire to places related to RV6.
- Nothing about SW2 elsewhere!

Unfortunately the magazine is missing a few pages and those appear to be related to the PCB for the compressor.  So whoever owned that magazine probably tore the pages out and built that unit.

Anyway, from all that it looks like the SW1, SW2, RV5, RV6 issue is resolved.

It looks to me like PR4 and RV5 just set the attack/decay time constants for the LED level display. They're nothing to do with the compressor rectifier, which is all down in the bottom-left corner, not the top-right. (I think...)

Actually the unit has two sections.   A compressor and a noise-gate.

The first OTA IC3a is the compressor and the rectifier is at the top right.   It also feeds the compression LEDs, which makes sense.   The compressor is a feedback type.

The second OTA IC3b is the noise gate and the rectifier is at the bottom left.   The Red and Green LEDs indicate the noise-gate status.

The other parts of the circuit are just for the input interfaces and output.

I am certain that the transistor on the top right is Q3.

Yes.  Also the diode near Q9.

also noticed that IC5c and IC6c have Pin9 as - and 10 as + (it should be backwards, right?  Positive is 9 and negative is 10 according to all the pinouts I have seen online).

Yes,  pin 9 is IN- and pin 10 is IN+

So the circuit *symbols* and connections are correct  but number labels for pin 9 and pin 10 are wrong on IC4c and IC6c.    IC5c looks OK.

The way I understand IC5c is it is wired as a Schmitt trigger.   The threshold is set by the pot RV6 by a negative voltage.  When the negative peaks from the rectifier IC5d get large enough (ie. negative enough) the output of IC5c will go high.  That at least matches the waveform picture.   
[It would be worth cross-checking that on the layout.]

[jfrabat]

OK, 4 important things;

IMPORTANT THING 1:

I had a LOOOOONG A$$ answer all typed up, and then Rob goes and posts his answer (while I was previewing mine!), which makes my post completely obsolete!  All that time typing, editing pictures to show what I was talking about, everything was wasted!  LOL!

IMPORTANT THING 2:

I'm now fairly convinced there's no SW2.  The connections for SW2 are actually wired connections for RV6. 
The reason for that are:
- RV6 has no pads on the PCB.
- The 'a' 'b' 'c' points seem to wire to places related to RV6.
- Nothing about SW2 elsewhere!

Rob, is it not possible that by NOT having RV6 connected (using the push/pull pot), then the gate is bypassed?  I do not yet understand enough to be able to tell, but could be the reason he used the DPSP Pot there...  If so I can intercept the cables from the board to the pot for the bypass switch.  This is not so critical, because I can always adjust the depth of the gate to make it negligible, but just wondering.

IMPORTANT THING 3

How about the transistor swap?  I cannot find BC212, but 2N3906 seems fairly close in specs.  What do you think?



IMPORTANT THING 4

EDIT: There goes Rob again editing his answer while I answer, making this point obsolete!  HAHAHA!

Thanks to all for your help!!!

[Rob Strand]

All that time typing, editing pictures to show what I was talking about, everything was wasted!  LOL!

It happens to everyone.  Especially long posts because people can probably post 5 short posts in the same time.
Sometimes I just leave the redundant post there.

Rob, is it not possible that by NOT having RV6 connected (using the push/pull pot), then the gate is bypassed?  I do not yet understand enough to be able to tell, but could be the reason he used the DPSP Pot there...  If so I can intercept the cables from the board to the pot for the bypass switch.  This is not so critical, because I can always adjust the depth of the gate to make it negligible, but just wondering.

It's certainly possible.

The main point that makes me think that there is no SW2, and the 'a' 'b' 'c' connections are for pot, is there is no PCB pads for RV6 on the layout.    You can only see trace lines and there is no room for any PCB pads.   The other pots have clearly visible PCB pads.

The evidence that SW2 exists is the part list.

So one possibility would be the switch disables the noise gate function.   That would make a lot of sense.

Putting all that together, the switch is wired directly to the pot and *then* the 'a', 'b', 'c' wires connect to the PCB.

How about the transistor swap?  I cannot find BC212, but 2N3906 seems fairly close in specs.  What do you think?

It should work fine.     (FYI, IIRC the transistor pinouts are different.)

EDIT: There goes Rob again editing his answer while I answer, making this point obsolete!  HAHAHA!

Sorry I often do that.   One reason is I've lost posts when my browser crashes.  It's a real pain if the post is long or I had to sift through a lot of details.   The other reason is I hate reading posts in the tiny edit window.   Sometimes Preview hangs so I just prefer to post.

[Rob Strand]

Forgot to mention, it's also odd there's notSW2 on the schematic and nothing in the text.  I'll have to read over the article again and see if there is any subtle hint about disabling the noise-gate.

[jfrabat]

Forgot to mention, it's also odd there's notSW2 on the schematic and nothing in the text.  I'll have to read over the article again and see if there is any subtle hint about disabling the noise-gate.

They mention it in the HOW IT WORKS sidebar (middle of the second paragraph).

[jfrabat]

Just realized the 2N3904 is a NPN, while the BC212 is a PNP...  The search continues for a replacement!

[Rob Strand]

They mention it in the HOW IT WORKS sidebar (middle of the second paragraph).

You did well to find that.    I just realized the PDF search doesn't find anything in the "HOW IT WORKS" section, perhaps because the page is rotated.

OK that's progress.   We know SW2 exists and what it does.   

The question now is how is it wired.

The "HOW IT WORKS" text says:
"... the signal passes through the noise gate VCA constructed around the other half of IC3 (unless this is bypassed by SW2) and thence to the outputs ..."

If I read that literally I would think the switch completely bypasses the noise gate.   From the way the schematic is drawn it doesn't seem likely that IC2c pin 7 is switched to R27, C14, C15 via SW2.    However, I suppose it's possible if they added the bypass after the schematic was done.

I tried to trace the circuit from the PCB overlay in fig 9.   Too many tracks go to the underside of the board so I cannot work it out.    The idea was to see if R27, C14, C15 are all wired via PCB tracks back to IC3b pin 9 + R26.

Can you see R27 on layout in fig 9?  I can't see it.  Some of the part labels are smudged.

The other option is to disable the noise gate by forcing IC5c on.  With the wires shown that would probably mean R71 is switched to ground to disable the noise-gate, or, R71 is switched to the wiper of RV6 to enable it.

[Rob Strand]

Just realized the 2N3904 is a NPN, while the BC212 is a PNP...  The search continues for a replacement!

2N3906 is the PNP version.

I must have tired eyes today my brain is filling in the blanks.  Not good for checking stuff.

[jfrabat]

So one possibility would be the switch disables the noise gate function.   That would make a lot of sense.

Putting all that together, the switch is wired directly to the pot and *then* the 'a', 'b', 'c' wires connect to the PCB.

So if I understand correctly, when the switch is "open", the 'a', 'b', 'c' wires dont connect to anything, and the gate is bypassed.  But with the switch closed, the 'a', 'b', 'c' wires connect to the PCB and enable the gate.  That was my understanding when I pointed out that maybe the switch could work without being connected, but I am not sure if it is what you meant.  Am i correct?

[Rob Strand]

So if I understand correctly, when the switch is "open", the 'a', 'b', 'c' wires dont connect to anything, and the gate is bypassed.  But with the switch closed, the 'a', 'b', 'c' wires connect to the PCB and enable the gate.  That was my understanding when I pointed out that maybe the switch could work without being connected, but I am not sure if it is what you meant.  Am i correct?

You only really need to leave R71 open.  However if IC5c pin 8 went negative the circuit would get stuck with the noise gate disabling the signal.  That's why I was thinking R71 needs to be grounded (or positive).    If you have an SPDT switch the common connects to R71 then the switch selects ground one one contact or the wiper of RV6 on the other.

If the switch disconnected the pot from R70 it should work like grounding R71.  As R71 would get grounded through the pot.

[jfrabat]

If you have an SPDT switch the common connects to R71 then the switch selects ground one one contact or the wiper of RV6 on the other.

If the switch disconnected the pot from R70 it should work like grounding R71.  As R71 would get grounded through the pot.

OK, I will do it this way!  Ordered all missing parts and power supply.

[Rob Strand]

OK, I will do it this way!  Ordered all missing parts and power supply.

It's about as good as you can do.

I tried to find another copy of the article with clearer photos, clearer layout and the PCB artwork but didn't get anywhere.

[ElectricDruid]

It looks to me like PR4 and RV5 just set the attack/decay time constants for the LED level display. They're nothing to do with the compressor rectifier, which is all down in the bottom-left corner, not the top-right. (I think...)

Actually the unit has two sections.   A compressor and a noise-gate.

The first OTA IC3a is the compressor and the rectifier is at the top right.   It also feeds the compression LEDs, which makes sense.   The compressor is a feedback type.

The second OTA IC3b is the noise gate and the rectifier is at the bottom left.   The Red and Green LEDs indicate the noise-gate status.

The other parts of the circuit are just for the input interfaces and output.

Ah, ok, thanks Rob. That's clearer now.